TABLE 2.
List of differentially expressed genes with concordant cis associations across populations
| Quantiles (rank)
|
MAFb
|
cis eQTL P-values
|
||||||
|---|---|---|---|---|---|---|---|---|
| Gene name | t-test | Batch corrected | ICE | rsIDb of cis SNP | CEU | ASI | CEU | ASI |
| TECA1 | 20.84% (840) | 30.45% (1227) | 0.55% (22) | rs11996666 | 0.12 | 0.67 | 3 × 10−7 | 2 × 10−10 |
| UGT2B17a | 1.44% (58) | 7.25% (292) | 0.02% (1) | rs3100645 | 0.68 | 0.15 | 1 × 10−11 | 9 × 10−37 |
| CSTBa | 13.87% (559) | 6.87% (277) | 1.19% (48) | rs4818868 | 0.38 | 0.82 | 2 × 10−8 | 9 × 10−10 |
| DFNA5 | 7.05%c (284) | 2.38%c (96) | 2.63% (106) | rs754555 | 0.13 | 0.49 | 9 × 10−8 | 1 × 10−13 |
| PEX6a | 21.49% (866) | 19.18% (773) | 0.84% (34) | rs2296804 | 0.59 | 0.36 | 2 × 10−15 | 8 × 10−11 |
| POMZP3a | 23.40% (943) | 10.77% (434) | 0.15% (6) | rs17718068 | 0.29 | 0.06 | 3 × 10−22 | 1 × 10−19 |
| RABGGTA | 49.95% (2013) | 24.44% (985) | 3.77% (152) | rs3940231 | 0.35 | 0.58 | 2 × 10−9 | 4 × 10−17 |
| RRM1 | 15.71% (633) | 12.08% (487) | 5.76% (232) | rs10500594 | 0.91 | 0.75 | 1 × 10−8 | 1 × 10−18 |
| AP3S2a | 6.50% (262) | 5.83% (235) | 10.27% (414) | rs9920421 | 0.31 | 0.16 | 8 × 10−9 | 2 × 10−6 |
| NUBP2a | 10.32% (416) | 11.22% (452) | 6.43% (259) | rs344359 | 0.18 | 0.32 | 3 × 10−9 | 3 × 10−11 |
| DNAJC15a | 19.13% (771) | 10.05% (405) | 3.37% (136) | rs9594865 | 0.76 | 0.90 | 2 × 10−9 | 1 × 10−11 |
The boldface entries have (1) cis associations with P < 2.5 × 10−8 in at least one population, (2) cis associations with P < 10−5 at the same SNP in the other population, (3) the difference of minor allele frequency (MAF) between populations >0.1, and (4) the strength of differential expressions between populations ranked among the top 10% of 4030 genes, using three different methods. Only 2 and 3 genes pass the criteria using the standard t-test and t-test after batch correction, respectively. Ten genes are identified as significant using the ICE eQTL approach, including four previously unreported differentially expressed genes. The genes are ordered by the difference of MAF between populations.
Previously identified by Spielman et al. (2007).
If multiple methods identify the same gene with a different strongest SNP, the SNP is selected according to the method providing the strongest rank in differential expression.
Although two methods have ranked the gene as among the top 10% in terms of differential expressions, no cis associations pass the P-value threshold described above.